The long-term goal of the proposed study is to better understand the determinants and consequences of college drinking. High prevalence of binge drinking and alcohol-related problems in college students has been well documented in a number of epidemiological studies. However, critical limitations of many studies such as cross-sectional assessment, lack of matched non-college peers, and failure to control for genetic and family background factors precluded the testing of causal models. The proposed study will addressthese limitations by implementing a longitudinal, prospective, and genetically informative design, in which the first assessment takes place prior to college, and family-based non-collegiate controls (twins and siblings) are used. Using cohort-sequential sampling, three cohorts of twins born in 1984-987 will be ascertained from birth records (n=2300), and assessed using telephone diagnostic interviews and a mailed questionnaire. The first assessment will take place in the senior high school year. Follow-up assessments of both collegiate and non-collegiate twins will be conducted in the first and fourth years post-high school, with brief assessments in years 2 and 3. In addition, 516 siblings of twins in their high school and college years will be ascertained. Assessments will include alcohol consumption measures, DSM-IV alcohol abuse and dependence symptoms, and risk factors such as family history of alcoholism, behavioral undercontrol, and personality traits. Specific aims of the study are: (1) to assess prospectively the effects of transition to college and, alternatively, entry into workforce on changes in drinking behavior; (2) to determine, using family-based non-collegiate controls, the relative role of dispositional (pre-existing) factors and high-risk college environment in the etiology of problem drinking in college students; (3) to test, using longitudinal genetic design, competing models of causal relationships between genetic factors, family environment, personality, psychopathology, pre-college alcohol involvement, and college drinking and alcohol-related problems. Data analysis will use structural equation modeling techniques and emphasize the testing of causal models such as mediational and risk-modifier models. This study will provide a greater understanding of the interplay of genetic and high-risk environmental factors in college drinking and can inform prevention and intervention strategies on college campuses.